OpenCV Color Calibration

A Python script for color correction and calibration using OpenCV and a ColorChecker chart. This tool detects a ColorChecker in an image, calibrates the colors, and applies the calibration to images or live video streams.

Table of Contents

• Prerequisites
• Installation
• Usage
  • Calibrate an Image
  • Apply Calibration to an Image
  • Apply Calibration to Live Camera Feed
• Code Structure
• Contributing
• License

Prerequisites

• Python 3.x
• A ColorChecker chart (Macbeth ColorChecker)
• OpenCV with extra modules (opencv-contrib-python)

Installation

First, uninstall any existing opencv-python package:

python -m pip uninstall opencv-python

Then, install the opencv-contrib-python package:

python -m pip install opencv-contrib-python==4.10.0.84

This version includes the cv2.ccm module required for color calibration.

Usage

Calibrate an Image

1. Place a ColorChecker chart in your scene and capture an image.

2. Run the script to detect the ColorChecker and calibrate the image:

   python your_script.py

   In the script, ensure the main function is called with the path to your image:

   if __name__ == "__main__":
       main(image_path='your_image.jpg')

3. The script will:

   • Detect the ColorChecker in the image.
   • Save the calibration parameters to calibration_params.pickle.
   • Calibrate the image and save it as calibrated_output.jpg.
   • Display the original and calibrated images.

Apply Calibration to an Image

To apply the saved calibration to another image:

1. Ensure you have the calibration_params.pickle file generated from the previous step.

2. Run the script with the calibration file and the image to calibrate:

   python your_script.py

   Modify the main function:

   if __name__ == "__main__":
       main(calibration_file='calibration_params.pickle', calibrate_image='image_to_calibrate.jpg')

3. The script will:

   • Load the calibration parameters.
   • Apply the calibration to the specified image.
   • Save the calibrated image with a prefixed name (e.g., calibrated_image_to_calibrate.jpg).
   • Display the original and calibrated images.

Apply Calibration to Live Camera Feed

To apply the calibration to a live video stream from your camera:

1. Ensure you have the calibration_params.pickle file.

2. Run the script with the calibration file:

   python your_script.py

   Modify the main function:

   if __name__ == "__main__":
       main(calibration_file='calibration_params.pickle')

3. The script will:

   • Load the calibration parameters.
   • Open a video stream from your default camera.
   • Apply the calibration to each frame in real-time.
   • Display the original and calibrated video streams.
   • Press q to exit the video stream.

Code Structure

• detect_color_checker(image): Detects the ColorChecker chart in the image and extracts color patches.
• calibrate_image(image, color_patches): Calibrates the image using the extracted color patches.
• apply_color_correction(image, model): Applies the color correction model to an image.
• save_calibration_params(color_patches, filename): Saves the calibration parameters to a file.
• load_calibration_params(filename): Loads calibration parameters from a file.
• reconstruct_model_from_params(params): Reconstructs the color correction model from saved parameters.
• calibrate_camera(calibration_file): Applies the calibration to a live video stream.
• main(): Orchestrates the calibration and application processes based on provided arguments.

Contributing

Contributions are welcome! Please open an issue or submit a pull request for any improvements or bug fixes.

License

This project is licensed under the MIT License - see the LICENSE file for details.

---